The present invention relates to a method of and an apparatus for reproducing audio information, particularly to an audio information reproducing method and an audio information reproducing apparatus, which are all suitable for speed-variable reproduction of an audio source as well as for changing a reproducing speed in a real time reproduction.
When an audio signal is to be reproduced slowly or quickly, the audio data is usually extended or compressed in the direction of a time axis. Namely, when an audio signal is to be reproduced slowly, the time is extended. On the other hand, when an audio signal is to be reproduced quickly, the time is compressed. In this way, when the reproduction is made slow, the tone of a sound becomes low. On the other hand, when the reproduction is made quick, the tone of sound becomes high.
In order to stabilize the tone of a sound so as to keep it at a constant value, the wave form of an original signal may be partially cut out and extended or compressed in the direction of a time axis, and then connected together so as to make constant an entire time length. However, since discontinuous signals are to be connected together in such a process, some sharp pulses will occur at some joint positions.
A procedure for carrying out the above process is shown in FIG. 6. Namely, segments of the wave form of an original signal are cut out, compressed or extended in the time axis, and further attenuated to smooth both ends of each signal by applying window functions. In detail, since each window function itself presents a mountain-like waveform and is thus symmetrical left and right (with both sides of each function signal being attenuated), a resulting waveform obtained by applying the window functions to the original signal will also be smoothly attenuated on both sides thereof. Therefore, if segments of the resulting signal are each retarded a little when being overlapped one upon another (so as to be connected together), it is possible to obtain a smooth and continuous signal not involving any seam joints.
However, in the above described prior art, when the segments of waveform are overlapped upon one another, it has been found that some modulation waves will occur and undesired low frequency wave will be generated. Further, it has been found that a smaller interval between adjacent overlapped segments will cause a larger modulation wave. Moreover, since the waveforms are overlapped upon one another, some mutually adjacent signals will be overlapped together, undesirably producing a sound involving an echo. Although this does not cause severe problem when an information data is music data, the clarity of the audio data will be deteriorated. On the other hand, if an interval between adjacent overlapped segments is too large, reverberation sound will be undesirably beard, thus causing a deterioration in the clarity of audio data.